Browse for Publications by:

Browse for Journal Articles based on DHS data by:

orange publication summary banner small

Document Type
Analytical Studies
Publication Topic(s)
HIV, HIV Prevalence, HIV/AIDS Knowledge, Attitudes, and Behavior, Nutrition
Burkina Faso, Cameroon, Ethiopia, Ghana, Guinea, Kenya, Lesotho, Malawi, Niger, Rwanda, Senegal, Zimbabwe
Sarah E.K. Bradley and Vinod Mishra and Macro International Inc. Calverton, Maryland, USA
Publication Date
September 2008
Publication ID

Order a Hard Copy: Please use electronic copies of DHS publications whenever possible. Hardcopies of publications are intended primarily for those in developing countries where internet connections are limited or unavailable.


Background In the absence of antiretroviral therapy (ART), HIV-positive people will lose weight and eventually become wasted or severely underweight. However, levels of obesity and overweight seem to be increasing in several sub-Saharan African countries severely impacted by HIV and with little or no ART available. In many of these countries, HIV prevalence is higher among overweight people than those of normal or below-normal body weight. Objectives To better understand this apparent anomaly, this study investigates the relationship between HIV, body mass index (BMI), and hemoglobin levels (biomarker for anemia) among adult women in sub-Saharan Africa. The study also investigates the relationship between HIV, women’s nutritional status, and their breastfeeding mode (exclusive, mixed, or none). Methods The data come from 12 nationally representative Demographic and Health Surveys (DHS) conducted between 2003 and 2006 in Burkina Faso, Cameroon, Ethiopia, Ghana, Guinea, Kenya, Lesotho, Malawi, Niger, Rwanda, Senegal, and Zimbabwe. Because height and weight were not measured for men in the majority of surveys, this study focuses only on women. Blood samples were collected for HIV and anemia testing, following internationally accepted ethical standards and procedures. HIV and anemia test results were linked anonymously to socio-demographic and height and weight data. The study used descriptive and multivariate statistical methods to examine the relationships between HIV status and BMI, hemoglobin concentration, and breastfeeding mode. In multivariate analysis, data from all 12 countries were pooled together and sampling weights were adjusted according to the relative population sizes of the countries. Results HIV and BMI. HIV prevalence is highest among overweight women in 5 of the 12 countries studied, and it is highest among underweight women only in Malawi. In seven countries, the percentage of HIV-negative women who are underweight is greater than the percentage of HIV-positive women who are underweight. Similarly, in half of the countries studied the percentage of HIV-positive women who are overweight is greater than the percentage of HIV-negative women who are overweight. In pooled analysis, HIV-positive vs. HIV-negative status is significantly positively associated with being overweight, and is significantly negatively associated with being underweight. After controlling for wealth and other socio-demographic characteristics, however, the relationship reverses direction, and HIV-positive status becomes significantly positively associated with being underweight. HIV and Anemia. HIV prevalence increases monotonically with severity of anemia for all countries except Ethiopia. In every country studied, the percentage of HIV-positive women who are anemic is much larger than the percentage of HIV-negative women who are anemic. In pooled analysis, HIV-positive status is significantly negatively associated with hemoglobin concentration, and the relationship grows stronger when background characteristics and BMI are controlled for. HIV, Nutrition, and Breastfeeding. Only limited data are available on women’s breastfeeding mode by HIV status, due to the small number of women who are HIV-positive and who have children less than six months old at the time of interview. In pooled multivariate models, however, there does not appear to be any negative impact of breastfeeding on the nutritional status of HIV-positive mothers compared with non-breastfeeding HIV-negative mothers. In fact, HIV-positive women who exclusively breastfeed their infants have slightly higher hemoglobin concentrations than HIV-negative women who do not breastfeed, after controlling for BMI and background characteristics. Conclusions HIV is independently associated with poorer nutritional status—particularly low hemoglobin concentration—after controlling for household wealth and other socio-demographic characteristics of the women studied. This study therefore provides evidence to support increasing food security and providing nutritional supplementation to HIV-positive people. At the same time, this study shows that in bivariate analysis HIV is concentrated among overweight women in several sub-Saharan African countries, emphasizing the fact that HIV does not affect only those who “look sick” or appear malnourished. Implications for the general public include both an emphasis for everyone to get tested regardless of how healthy they may look, and for individuals to remember to use protection with every partner even if they appear healthy. Counselors should encourage people to be tested for HIV regardless of their current BMI, and they should consider that clients could be HIV-positive even if they appear overweight. If past trends continue, the proportion of people in sub-Saharan Africa who are either overweight or underweight will grow. This dual burden has particular implications for ART, which is becoming more widely available as many countries work towards scale-up. To address potential problems with ART use in overweight and underweight patients, clinicians should consider their patient’s nutritional status, including titrating drug combinations to patient’s body weights when possible to limit adverse effects due to nutritional differences.